Documentos de Académico
Documentos de Profesional
Documentos de Cultura
1-1-2012
Recommended Citation
Thoene, Whitney Sue, "The Impact of Social Networking Sites on College Students' Consumption Patterns" (2012). Theses,
Dissertations and Capstones. Paper 245.
This Thesis is brought to you for free and open access by Marshall Digital Scholar. It has been accepted for inclusion in Theses, Dissertations and
Capstones by an authorized administrator of Marshall Digital Scholar. For more information, please contact zhangj@marshall.edu.
THE IMPACT OF SOCIAL NETWORKING SITES
ON COLLEGE STUDENTS CONSUMPTION PATTERNS
A thesis submitted to
The Graduate College of
Marshall University
In partial fulfillment of
the requirements for the degree
of Master of Arts
Communication Studies
by Whitney
Sue Thoene
Approved by
Marshall University
May 2012
Dedication
This thesis is dedicated to my beloved family. To my parents, Dale and Nancy,
both of you have loved and supported me throughout all my endeavors. To my brother,
Wesley, you are always a source of inspiration and encouragement. Family is important
ii
TABLE OF CONTENTS
Dedication...........................................................................................................................ii
Abstract................................................................................................................................v
Chapter 1: Introduction........................................................................................................1
Chapter 2: Literature Review.............................................................................................. 3
Consumption Patterns......................................................................................................6
Predictor of Sales.............................................................................................................9
Gender............................................................................................................................. 9
Social Media Usage Frequency..................................................................................... 11
Chapter 3: Methodology....................................................................................................13
Participants.................................................................................................................... 14
Materials and Procedure................................................................................................ 15
Chapter 4: Results..............................................................................................................17
The Sample....................................................................................................................17
Research Question #1.................................................................................................... 22
Facebook....................................................................................................................23
Twitter........................................................................................................................27
Research Question #2.................................................................................................... 33
Facebook....................................................................................................................38
Twitter........................................................................................................................40
Research Question #3.................................................................................................... 43
Facebook....................................................................................................................44
Twitter........................................................................................................................46
Regressions....................................................................................................................48
Chapter 5: Discussion........................................................................................................55
Research Question #1.................................................................................................... 55
Number of Friends or Followers............................................................................... 56
Number of Companies Friended or Followed........................................................... 57
Shopping Frequency..................................................................................................58
Frequency of Correspondence...................................................................................59
Frequency of Sales Information................................................................................ 60
Research Question #2.................................................................................................... 60
Research Question #3.................................................................................................... 64
Summary........................................................................................................................67
Limitations.....................................................................................................................68
Heuristics.......................................................................................................................68
Appendixes........................................................................................................................ 70
Appendix A: Questionnaire...........................................................................................71
Appendix B: Thesis Defense Executive Summary........................................................76
Appendix C: Thesis Defense PowerPoint..................................................................... 79
Works Cited....................................................................................................................... 92
iii
LIST OF TABLES
iv
Abstract
This research study examines the effect of social media, particularly Facebook
and Twitter, on the purchasing habits of college students by testing for correlations
research also examines the role of gender and social media usage frequency on
consumption patterns. Findings revealed that both Facebook and Twitter are being used
to obtain sales information and promotions. Furthermore, gender has an impact on both
social networking sites. Additionally, this study found the higher the frequency of social
media usage the more likely customers are to shop at the businesses they have
befriended.
v
1
Chapter 1: Introduction
Social media is a popular trend today, especially among college students.
Businesses are always looking for new ways to reach customers, especially ones readying
to enter the work force. This study examines the impact of social media on the buying
Several past research studies have focused on the effects of electronic word-of-
mouth communication on consumer behavior. Hu, Liu, & Zhang (2008) discovered a
positive relationship between products with good online reviews and the sales of that
product. The more a reviewer was exposed to the product and the more credibility the
reviewer had the more influential the individuals feedback became. This word-of-mouth
communication had an impact on potential buyers decisions. This study was augmented
by Hennig-Thurau, Gwinner, Wlash, and Gremlers 2004 study. Their research explored
belonging, potential payment, and a desire to aid others are a few of the reasons they
discovered.
Other research has focused on the role of social media. Smock, Ellison, Lampe,
and Wohn (2011) analyzed the reasons people use Facebook through the uses and
gratification model. Among the reasons they found were for social interaction,
Ogle, Tan and Reitz (2011) found that many social media websites are being utilized by
companies to communicate with customers. Customers are able to express their opinions
about a company or brand. LaDuque (2010) discovered that companies use social media
sites to communicate directly with the customer, as companies are able to increase brand
2
loyalty, create sales leads, and increase publicity through this medium. Furthermore,
social media sites are being used for marketing research by both companies and
The purpose of this study is to analyze the impact of social media, particularly
Facebook and Twitter, on students consumption patterns. This thesis also examines if the
role of social media on consumption patterns differs between genders. For example, are
males or females more likely to use promotions offered by companies they have befriended
or follow? The final research question seeks to determine whether or not word-of-mouth
recommendations made via social media platforms have varying impacts based on usage
frequency. This research question analyzes the number of times student check their social
media accounts and the frequency they use the promotions offered.
3
one of the most well-known social media sites is Facebook (Facebook, 2011). This site
currently boasts 800 million active users, and over 50% of active users log on to the site
every day (Facebook, 2011). Members are able to connect with friends on the site, and
the average user maintains approximately 130 friends (Facebook, 2011). Further, more
than 350 million of these users access the site through a mobile device (Facebook,
2011). According to Eldon (2011), 51.2% of users are male, whereas 48.8% are female.
In terms of age, 20.6% of users are between the ages of 13 and 17; 25.8% are between
the ages of 18 and 25; 26.1% are between the ages of 26-34; and 27.5% are over the age
of 35 (Eldon, 2011).
Many studies have supported the idea that a positive correlation exists between
online product reviews and sales (Hu, Liu, & Zhang, 2008). In 2008, Hu, Liu, and Zhang
tested the idea that products with favorable reviews sell better than products with
unfavorable reviews. In addition, Hu et al. sought to determine whether the quality of the
reviewer (reviewer appears more reliable) would influence the effectiveness of the
favorable and unfavorable online reviews. Finally, the researchers found that consumers
also consider reviewer quality and reviewer exposure (the frequency the name of a
review is present in the community) when evaluating an online review (Hu et al., 2008).
other factors (Hu et al., 2008). Managers who are seeking to improve electronic word-of-
individuals who used this communication medium. In 2009, Ross, Orr, Sisic, Arseneault,
Simmering, and Orr explored the personalities and motivations of individuals who used
the social media site Facebook. To this extent, the researchers applied the Five Factor
Model to Facebook usage. Ross et al. found individuals who exhibited higher levels of
extroversion were more likely to belong to more Facebook groups; however, this
personality trait was not associated with having a larger number of friends. The
Walsh, and Gremler defined eleven motivations to explain why consumers engage in
examine these motivations. Based on the results, four motivations were discovered:
social benefits, economic benefits, concern for others, and self-enhancement (Hennig-
Thurau et al., 2004). Social benefits refers to the idea that consumers gain a sense of
any payment a reviewer may receive (Hennig-Thurau et al., 2004). Concern for others
A similar study conducted by Hyllegard, Ogle, Yan, and Reitz (2011) sought to
brands on the social networking site. The researchers found students use the site to
establish personal connections with others and use the site to create affiliations with
brands that define who they are and help them establish a sense of self (Hyllegard et al.,
2011). Thus, these motives are similar to the social benefits motive discovered by
Casteleyn et al. (2009), and the desire for self-expression cited by Pempek, Yermolayeva,
and Calvert (2009). In addition, Hyllegard et al. (2011) found that students fanned
companies and brands to become market mavens who could receive and disseminate
consistent with the concern for others motive and self-enhancement motives
about a product (self-enhancement), and then share this information with friends
In 2011, Smock, Ellison, Lampe, and Wohn applied the uses and gratification
approach to analyze why individuals use Facebook in general, as well as why they use
certain functions on the website. The researchers found that users who update their status
are motivated chiefly by a desire for expressive information sharing, whereas individuals
interaction. However, individuals who posted on friends walls did so for professional
6
advancement, social interaction, and habitual pass time. Two motives, professional
advancement and social interaction, were discovered as underlying reasons why users
sent private messages. Smock et al. also found social interaction was the only significant
motive discovered in the usage of Facebooks chat feature. Finally, the usage of groups
on the site was positively influenced by expressive information sharing, and negatively
by social interaction.
Consumption Patterns
In addition, companies have also begun using social media sites as a way to better
communicate with customers. According to Hyllegard, Ogle, Yan, and Reitz (2011),
between 66%-96% of consumer goods companies have adopted social media, including
Facebook (p. 601). One feature of this site is the ability of users to fan particular
products or brands so that they receive information about these products. In addition, this
brand. Facebook reports the average user can express interest in over 900 million objects,
According to LaDuque (2010), companies can also utilize social media platforms to
create personalized experiences, to increase brand loyalty, to generate sales leads, and to
increase exposure. For example, companies which provide entertaining videos or valuable
incentives may have their offerings shared between social media users through electronic
Facebook allows individuals to like pages, meaning they will receive updates
7
regarding the product or brand they have liked. Even mutual fund firms have begun
using this medium to communicate with customers, create interactive scavenger hunts,
Social media can also be used by both companies and customers for marketing
research. In 2007, Casteleyn, Mottart, and Rutten (2009) sought to understand the
behaviors of individuals who joined groups on this social networking site. These groups
Casteleyn et al., market researchers could gain insight into individuals feelings about a
brand from reading wall posts. These wall posts could potentially include comments
about the brand or photographs involving the brand. These researchers hypothesized
researchers could better understand this consumer behavior by considering the agent (the
individual posting the information), the act (what he or she posted), agency (the method
used to post the information), the scene (the context of the post), and the purpose (why
the information was posted). These findings are of significant importance, as Pempek et
al. (2009) found that the average student belonged to 24.58 groups; however, Pempek et
al. also found that active participation within these groups was rare.
to collect information about companies and brands (Barnes, 2008). Barnes found 70%
like Facebook was deemed by customers to be of little value, respondents under the age
of 25 viewed information on social media sites more positively. However, the role of
8
brand based on online reviews (Barnes). Thus, positive reviews can improve sales,
Despite these potential benefits, companies must also be cautious when using
Facebook. First, culture has developed on the website, and companies must be careful to
adhere to the cultural norms preset on the site (Vorvoreanu, 2009). To understand how
conducted six focus groups with 35 college students. Vorvoreanu discovered some users
feel as though corporations do not belong on the site, as it was meant for friends to
interact. Consistent with other research studies, many users view their profiles as a
means of self-expression, and becoming fans of a company allows them to express their
maintained a presence on the site, though this was because they often knew the owners
begun maintaining a presence on Facebook, they also reported that they would be
interested in receiving discounts and gifts from these organizations through the social
In addition, social media traffic also exhibits a higher bounce-rate (85%) than
search engine traffic (50%), meaning people who access sites through social media are
Predictor of Sales
Traditionally, researchers have focused on the role of word-of-mouth
Gremler, 2004). However, the introduction of the internet has led to an increasing focus
media may have on shaping consumers opinions (Hennig-Thurau et al., 2004; Hu, Liu,
& Zhang, 2008). Much of the research in this area focuses on how brand managers can
use electronic word-of-mouth to increase brand recognition and improve sales (Hennig-
Gender
In a survey of 272 undergraduate and graduate students, Budden, Anthony,
Budden, and Jones (2007) found females spent more time on social networking sites
Facebook and MySpace than males; however, these differences were not found to be
reported by Eldon (2011); however, the differences found by Eldon were not only
reported and not tested; as a result, they may not be statistically significant. In addition,
any differences in findings may be due to the sample or the dates on which the data were
collected. Males were also found to spend significantly more time on YouTube than
females (Budden et al., 2007). Budden et al. (2007) also discovered that upperclassmen
spend more time accessing the internet than underclassmen and graduate students.
However, the results of the study indicate many college students still utilize traditional
The results of Budden et al. (2007) were later supported by the findings of
Gerlich, Browning, and Westermann (2010). In this study, 141 undergraduate students
were surveyed to determine media usage among college students. Consistent with
Gerlich et al., the researchers found no statistically significant differences between males
and females in the amount of time spent using the internet. In addition, no fundamental
differences were discovered between females and males in terms of social media usage
In 2008, Foregger used a mixed methods approach to apply the uses and
gratification theory to Facebook. For this study, Foregger sought to understand how
undergraduate college students used the social networking site, as well as the
gratifications they received from using it. Prior to the main study, Foregger conducted a
preliminary study by listing thirteen Facebook uses and asking 185 students to rate how
often they use Facebook for each of these purposes. Foregger then developed a survey
responses, Foregger determined students use Facebook for nine reasons: pass time,
connections, sexual attraction, utilities and upkeep, establish old ties, accumulation,
social comparison, channel use, and networking. In addition, the researcher found
females spend more time on the site than males, a trend which may explain why females
tend to have more Facebook friends on the site than males (Pempek et al., 2009). Female
freshmen, female sophomores, and female juniors spent the most time on Facebook,
while male juniors and male seniors spent the least amount of time on the site (Foregger,
2008).
11
students in particular used Facebook. In this study, 92 college students recorded their
Facebook usage in terms of time and the functions they used on the site. Respondents
were also asked why they used the site. Based on the findings of the study, respondents
reportedly averaged 27.93 minutes on the site on weekdays and 28.44 minutes per day on
weekends (Pempek et al.). The majority of students also used the site in the evenings as
the hours of 9:00PM to 12:00AM were the mostly commonly reported log-in times
(Pempek et al.). In terms of motivation, students used the site to communicate with
friends (Pempek et al.). The site was more often used by underclassmen, many of whom
used it to maintain contact with high school friends who attended other schools (Pempek
et al.). In addition, females had a larger number of friends than males on the site (Pempek
et al.). The researchers also discovered that students spend a greater amount of time
receiving content than posting it because they read the content posted by friends on the
site (Pempek et al.). Many respondents also reported that they used the website to
These findings supported the earlier results of Ellison, Steinfeld, and Lampe
(2007), who analyzed the relationship between Facebook and its effect on social capital.
To this extent, the researchers analyzed how Facebook (and similar sites) affected three
types of social capital: bridging social capital, bonding social capital, and maintained
social capital (Ellison et al., 2007). In this study, 286 students were surveyed and 94% of
them used Facebook. Ellison also discovered that these students spent between ten and
thirty minutes per day on the site, and these findings on daily usage are consistent with
those of Pempek et al. (2009). Ellison also found students use the site to maintain
12
existing relationships, and this finding is also consistent with the findings of Pempek et
al. (2009). Ellison et al. (2007) concluded positive relationships exist between Facebook
Much research has been done on social media. Companies are using social media
these sites have a positive impact on sales (Barnes, 2008). Furthermore, customers obtain
information about companies and brands on social media sites. This study analyzes the
research is inconsistent in gender usage of the Internet and social media sites. According
to Budden et al. (2007), females spend more time on social networking sites than males;
students. However, Gerlich et al. did not find any statistically significant differences
between gender usages of the Internet or social media. The research in this study
examines the differences, if any, between the genders and whether gender affects the
previous research has found that people use Facebook frequently and for self-expression
and communication with friends. This study investigates the relationship between social
Chapter 3: Methodology
In this study, a quantitative research survey was created to analyze the effect of
social media on the consumption patterns of students at collegiate institution. The survey
instrument focused on Facebook and Twitter. The research studys purpose was to
Facebook and Twitter on the purchasing habits of students. Therefore, the following
patterns?
consumption patterns.
consumption patterns.
between genders?
After compiling the data, several tests were run to find relationships between the
tests were used to examine the strength of relationships between variables; these tests were
conducted at a significance level of .05. To test research question 2, t-tests were run to test
for differences between genders at a significance level of .05. Pearson correlations were also
used to analyze research question 3, and the relationships between usage frequency and
other dependent variables were tested at a significance level of .05. Finally, analysis of
variance tests were conducted to determine whether any other independent variable relating
.05.
Participants
The research study was based on a convenience sample. There were 275 students
from a Midwestern university who participated in this study. The population of the study
was approximately 13,570 students (College Toolkit). Participants were recruited both
online and in person. Several faculty members announced the study in their classes.
15
electronic survey was opened. Facebook posts were published and e-mails were sent to
and flyers were distributed on campus as well. Students were aware participation was
voluntary.
Before beginning the survey instrument, respondents were first routed to the
Informed Consent Form. Once the form had been read and their age was verified,
respondents were able to access the survey instrument. No identifying information was
requested on the questionnaire nor was any identifying information recorded. The survey
was comprised of 28 questions (Appendix A), several of which were grouped together.
The first two questions pertained to the number and type of social accounts the
respondent maintains. The next eight questions asked about participants Facebook
accounts, whereas the following eight inquired about their Twitter accounts. The
subsequent two questions asked about their shopping habits; finally, the last eight
Respondents were initially asked how many social networking accounts they
maintained and then were asked to identify them. If the respondent did not have any
accounts, they were able to skip to the shopping preference set of questions; however, if
they confirmed that they had a Facebook account, they continued. The next question
inquired how frequently the respondent checked his/her Facebook account. The choices
ranged from 1 to 5; 1 was never, 5 represented multiple times a day. The next question
asked for the number of friends the respondent has on his/her account. Then, participants
following questions were clustered together. These four questions pertained to the
correspondence from these companies via Facebook, obtain information about sales,
specials, or coupons from these companies, and utilize these sales, specials, or coupons.
These were measured on a scale; 1 represented never and 5 denoted very frequently.
If respondents confirmed that they had a Twitter account, they were directed to
the Twitter portion of the survey; otherwise, they were able to skip to the set of questions
on shopping habits. The first two Twitter questions pertained to the frequency the
respondents checked their accounts and how many followers they have. The next groups
pertaining to the frequency that respondents shop at the companies they follow, receive
correspondence from these companies via Twitter, obtain information about sales,
specials, or coupons from these companies, and utilize these sales, specials, or coupons.
These were measured on a scale; 1 represented never and 5 denoted very frequently.
Next, participants were asked how often they shop online, then in person. These
inquiries operated on the same scale as the previous set. Finally, demographic questions
Chapter 4: Results
The Sample
survey was distributed at a Midwestern college. A total of 275 students took the survey.
Of these students, 82 were male, 158 female, and 35 did not provide a response. Thus, of
those who provided a valid response, 34.2% were male and 65.8% were female. These
Table 1
Gender of Sample
The gender demographics found in the survey are proportional with the
demographics of the overall college. A 2008 survey of students at the university found
that 44.8% of undergraduate students at the college were male, while 55.5% of
students at the college were female (College Toolkit). Thus, the sample was comprised
of more females than males, consistent with the population at the college.
Table 2
sophomores. In addition, 10.5% of respondents were juniors, 8.7% were seniors and
21.8% were graduate students. Notably, almost one-fifth (19.6%) of respondents did not
provide a class rank for the survey. When these nonresponses are excluded, 31.7% of
respondents were freshmen, 17.2% were sophomores, 13.1% were juniors, 10.9% were
seniors, and 27.1% were graduate students. These statistics are consistent with the
overall population at the participating university, as a 2008 survey of students found that
68.5% of students were undergraduate students, 29.3% were graduate students, and
A wide range of age groups were also present in the study. The age demographics
Table 3
Age of Sample
As shown in Table 3, almost half of the respondents were between the ages or 18
and 21. This result is not surprising, as a majority of traditional college students are
between these ages. In addition, 36 students (13.1%) did not provide an age in the
survey.
asked whether they were full-time or part-time students. A large percentage of students
(66.5%) were full-time students, as shown in Table 4. Of the people who responded,
83.6% of respondents were full-time. Based on this characteristic, the sample represents
the population, as a 2008 survey of students at the institution found that 87.5% of
undergraduate students were full-time, while 12.4% were part-time students (College
Toolkit).
20
Table 4
Respondents were also asked about their current employment status. Participants
were asked to select among three options: full-time, part-time, and no current
Table 5
provided a response, 29.3% held a full-time job, while 36.4% held a part-time job and
Respondents were also asked to report their marital status, and the results of this
Table 6
reported being single. Conversely, 17.8% of respondents were married and 13.1% did
The ethnicity of respondents was also asked in the survey, and responses to this
Table 7
Ethnicity of Sample
78.9% of respondents were of this ethnicity. Hispanics, Native Americans, and Asian
Americans had very little representation in the sample, as these three groups comprised
22
only 2.9% of the sample. African-Americans, meanwhile, comprised 4.4% of the sample.
In addition, 13.1% of respondents chose not to report their ethnicity. When these
nonresponses are eliminated from the analysis, the percentage of whites in the sample
students from this ethnic group (College Toolkit). A 2008 survey of the academic
institutions students found that 88.35% of students were Caucasian, whereas 5.66% were
Americans, Asian Americans, and Native Americans comprised 2.59% of the student
Research Question #1
The first research question sought to determine whether social media
websites like Facebook and Twitter affect consumption patterns. To examine this
research question, the following research question and hypotheses were established:
on consumption patterns.
23
on consumption patterns.
usage frequency of the sales, specials, or coupons received from these sites. This
information was collected for two social networking sites: Facebook and Twitter.
Because these scales represented interval-level data, Pearson correlations were used to
test these relationships; furthermore, these correlations were conducted using a .05
standard of significance.
The results of the Pearson correlations can be found in Table 8. Correlations with
number of Facebook friends was analyzed with the number of companies liked. This
test revealed a .140 coefficient of correlation, and this result was statistically significant
at a .05 level of significance. Thus, although statistically significant, the small correlation
coefficient indicates this relationship was weak; consequently, it can be argued there is
only a slight relationship between the number of friends an individual has on Facebook
and the number of companies they have liked. The relationship between the number of
friends
24
Table 8
respondents have on Facebook and the frequency at which they shop at companies they
have friended on Facebook was also examined. This test revealed a correlation
coefficient of .114; however, this result was not significant at the .05 level.
Likewise, a Pearson correlation was used to test the relationships between number
of Facebook friends and the amount of correspondence received from companies. This
test revealed a non-significant correlation coefficient of -.002, which indicates there was
no association between these two variables. Respondents were also asked how often they
Facebook, and a Pearson correlation was used to test the relationship between this
of .078 was found for these two variables. Finally, the relationship between usage of
these promotions and number of Facebook friends was tested using a Pearson correlation.
This test revealed a correlation coefficient of .114, though it was not statistically
significant.
First, correlations with the number of companies liked on Facebook and the
shopping frequency at these stores was analyzed. A correlation coefficient of .487 was
calculated and found to be significant at a significance level of .05. Thus, this positive
relationship was found to be moderate and substantial. Individuals who like companies
on Facebook are likely to shop at these stores. Next, the relationship between number of
companies. A correlation coefficient of .389 was discovered. This coefficient was found
definitive but small. Thus, as individuals like more companies on Facebook, they are
Next, the relationship between the number of companies liked on Facebook and
the amount of information received about sales and promotions was analyzed. A
that as individuals like more companies on Facebook, they are more likely to receive
more information on sales and promotions. Finally, the relationship between the number
of companies liked and how often sales and promotions are received through this channel
moderate, positive relationship. Thus, individuals who like more companies on Facebook
are more likely to take advantage of the sales and promotions being offered through this
medium.
Correlations with shopping frequency at stores liked on Facebook was the next
variable analyzed. First, the relationship between this variable and the amount of
these stores. In addition, the relationship between shopping frequency at stores liked on
Facebook and the amount of information received from companies about sales was
Facebook, they are more likely to take advantage of promotional offers received on this
site.
27
from companies on Facebook was analyzed in relation to the amount of sales and
promotional offers and the usage of sales and promotional offers. First, a correlation
between Facebook correspondence frequency and amount of sales and promotional offers
was moderate-to-strong and positive. Thus, the more correspondence individuals receive,
the more likely this information is to include sales and other promotions. Finally, a
correlation between correspondence frequency and usage of sales and promotions was
Finally, the relationship between the frequency at which information on sales and
promotions was received and the frequency at which these offers are taken advantage of
was analyzed. A Pearson correlation coefficient of .628 was found, which was
Twitter
Next, the relationships between Twitter variables were analyzed. The results of
Table 9
Correlations
twittercorresponden
twitterfollowers twittercompanies twittershopfreq ce twittersalesinfo twittersalesuse
twitterfollowers Pearson Correlation 1.000 .367** .505** .458** .537** .423**
Sig. (2-tailed) .000 .000 .000 .000 .000
N 176 174 173 171 173 171
In order to test the relationships between variables for Twitter, the first relationship
which was analyzed was the correlation between number of followers on Twitter and
number of companies followed. Because the scales represented interval-level data, this
relationship was tested using a Pearson correlation at a .05 level of significance. This test
revealed a correlation of .367 between these two variables, and this correlation was
definitive, yet small, meaning that as individuals follow more individuals they are slightly
more likely to follow more companies. The relationship between the number of followers
that respondents have on Twitter and the frequency at which they shop at companies they
.05 level of significance. This test revealed a correlation coefficient of .505, and this
result was found to be significant at the .05 significance level. This moderate correlation
indicates individuals who are more active on Twitter (as based on the number of
followers they have) are somewhat more likely to shop at companies they have followed.
Twitter and the frequency with which respondents receive correspondence from
correlation test was also used to measure the relationship between number of Twitter
followers and the frequency at which respondents receive information on sales, specials,
which was a positive, moderate relationship. Thus, individuals who have more followers
are more likely to receive information from companies about special deals. Finally, the
30
relationship between number of followers and the frequency at which respondents take
advantage of deals and specials from this medium was tested. A Pearson correlation test
individuals who have more followers are slightly more likely to use the sales offered
through Twitter.
Next, the relationship between the number of companies followed on Twitter and
other variables were tested. First, a Pearson correlation at a significance level of .05 was
used to test the relationship between number of companies followed on Twitter and how
often individuals shop at these companies they have followed. A correlation coefficient of
.523 was calculated, and this coefficient was found to be statistically significant. Thus,
there is a moderate, substantial relationship between these two variables: individuals who
follow more companies on Twitter are only slightly more likely to shop at these stores.
Twitter and the frequency at which they shop at these establishments was calculated. To
test this relationship, a Pearson correlation was used with an established significance
meaning the relationship was positive and substantial. Thus, individuals who follow more
companies are only slightly more likely to receive more correspondence from these
firms. These results seem to suggest that some companies may not be using Twitter to
correspond with their customers and thus are not using this channel to its fullest extent.
This theory is further supported by the next test which was conducted. In this test, the
relationship between the number of companies followed on Twitter and the frequency
with which customers received sales or promotional offers from these companies was
31
examined. Because the scale was interval-level, a Pearson correlation was used at a
significance level of .05. A coefficient of .582 was discovered, meaning the positive
companies, they are only slightly more likely to receive more correspondence from these
firms.
on Twitter and the frequency with which respondents used the sales information was
examined. This relationship was also tested using a Pearson correlation at a significance
level of .05. A .484 correlation coefficient was discovered, meaning the positive
relationship was substantial and moderate. Thus, individuals who follow more
companies on Twitter are only slightly more likely to take advantage of the offers
Next, the relationships between shopping frequency at companies that have been
followed and the other variables were examined. First, the relationship between how often
respondents shopped at companies they have followed and how often they received
correspondence from these companies was examined. A Pearson correlation test at a .05
correspondence from companies are more likely to shop at these stores. A Pearson
correlation at a .05 level of significance was also used to test the relationship between how
often respondents shop at stores they have followed on Twitter and how often they receive
information on sales from these locations. This test revealed a positive, dependable
individuals who receive sales information from these stores on Twitter are more likely to
Last, the relationship between how often respondents shopped at companies they
have followed on Twitter and how often they used sales information received through
this medium was explored. A Pearson correlation at a significance level of .05 was used
to test this relationship, and a .792 coefficient was calculated. Thus, the relationship was
deemed to be positive and marked. Individuals who often shop at stores they have
followed on Twitter are more likely to take advantage of sales they find on this site.
from companies on Twitter and other variables was examined. First, a Pearson
correlation was used to test the relationship between how often respondents received
correspondence from companies on Twitter and how often they receive information on
This correlation indicates that individuals who receive a great deal of correspondence
from companies are more likely to receive information on sales and coupons. This result
is logical as much of the correspondence received from companies is likely about sales.
from companies they follow and how often they take advantage of sales was examined.
A Pearson correlation test was used at a significance level of .05, and a correlation
coefficient of .870 was discovered. This coefficient represents a dependable and positive
33
relationship between the two variables. Thus, individuals who receive more
correspondence from companies on Twitter are more likely to take advantage of sales.
Last, the relationship between how often individuals receive sales information on
Twitter and how often they take advantage of these sales was examined. A Pearson
correlation was used to test the relationship between these two variables at a significance
strong, dependable relationship exists between the two variables. Thus, individuals who
receive more information on sales through Twitter are more likely to take advantage of
these sales.
Research Question #2
The second research question sought to determine whether the effectiveness of
To test this research question and these hypotheses, the mean responses for both
through 10), Twitter followers (11 through 12), and Twitter correspondence (14 through
34
17) were calculated; subsequently, t-tests for independent samples were utilized to
level of .05 Table 10 shows the mean response scores for females and males for each
Table 10
Group Statistics
Group Statistics
gender N Mean Std. Deviation Std. Error Mean
Table 11
Equal variances not -2.263 163.409 .025 -.415 .184 -.778 -.053
assumed
fbsalesuse Equal variances 10.213 .002 -2.980 219 .003 -.491 .165 -.815 -.166
assumed
Equal variances not -3.257 193.220 .001 -.491 .151 -.788 -.194
assumed
twittercheck Equal variances .126 .723 -.262 211 .794 -.061 .232 -.517 .396
assumed
Equal variances not -.262 152.983 .794 -.061 .231 -.517 .396
assumed
twitterfollowers Equal variances .008 .930 .044 164 .965 .010 .217 -.418 .437
assumed
Equal variances not .044 127.809 .965 .010 .217 -.420 .439
assumed
twittercompanies Equal variances .495 .483 1.214 199 .226 .207 .170 -.129 .543
assumed
Equal variances not 1.164 123.180 .247 .207 .178 -.145 .559
assumed
twittershopfreq Equal variances .030 .864 .008 198 .994 .001 .145 -.285 .287
assumed
Equal variances not .008 145.306 .994 .001 .144 -.283 .285
assumed
twittercorrespondence Equal variances .012 .912 -.014 197 .988 -.002 .131 -.260 .256
assumed
Equal variances not -.014 128.641 .989 -.002 .134 -.268 .264
assumed
twittersalesinfo Equal variances 1.688 .195 -.545 198 .586 -.079 .145 -.365 .207
assumed
Equal variances not -.568 158.226 .571 -.079 .139 -.354 .196
assumed
twittersalesuse Equal variances .200 .655 -.209 196 .835 -.028 .136 -.296 .240
assumed
Equal variances not -.211 147.075 .833 -.028 .134 -.294 .237
assumed
38
The first set of comparisons between females and males responses tested
differences in how often females and males checked their Facebook accounts. This
recommendations from their peers. As shown in Table 11, the first unit of output
represented a Levenes Test for Equality of Variance, which tested whether the two
groups (in this case, genders) displayed equal variance on the dependent variable (the
frequency at which they check their Facebook accounts). Because this output was .013,
and this statistic was less than the stated significance level of .05, one can assume the
variances are not equal. As a result, the bottom line of the output was used in this
analysis, and thus the calculated p-value was .176. Because the calculated p-value (.176)
was greater than the desired significance level of .05, the differences between genders on
The second variable which was tested related to the average number of friends
students have on the social networking site Facebook. As students have larger numbers
of friends on this site, the more likely they are to receive word-of-mouth
recommendations from these peers. The results of this test can be found in Table 11.
The Levenes Test for Equality of Variance determined a calculated p-value of .122.
Because this was larger than the desired significance level of .05, the variances were
equal between the two means, and thus the top row of output was used in the analysis.
Consequently, the calculated p-value (.401) was larger than the desired level of
39
significance (.05); thus, the differences between genders on this variable were not
The third variable which was tested related to the average number of companies
males and females like on Facebook, and the results of this analysis are shown in Table
11. The Levenes Test for Equality of Variance determined a calculated p-value of .547.
Because this was larger than the desired significance level of .05, the variances between
the two means were deemed to be equal. The calculated p-value (.505) was larger than
the desired level of significance (.05), and so the differences between genders in terms of
number of companies they have liked on Facebook were not statistically significant.
Next, the frequency with which students shop at stores they have liked on
Facebook was analyzed. As shown in Table 11, the Levenes Test for Equality of
Variance found a calculated p-value of .818, which was much larger than the desired
level of significance of .05. As a result, the variances between the two means were
determined to be equal. The corresponding calculated p-value of .240 was larger than
.05; consequently, no statistically significant differences were found between males and
females in regard to how often they shop at stores they have liked on Facebook.
The next column in Table 11 shows the results of a t-test for independent samples
Levenes Test for Equality of Variance showed the variances were equal, as the calculated p-
value of .377 was larger than the desired significance level of .05. As a result, the output on
the top level was used, and the calculated p-value of .296 was larger than the desired
be statistically significant.
Respondents were next asked how often they receive information on sales from
the companies they have liked on Facebook. As shown in Table 11, a Levenes Test
for Equality of Variance revealed a calculated p-value of .356, thus determining the
variances were equal. Based on this analysis, the calculated p-value was .028. Because
this p-value was less than the desired significance level of .05, the differences between
males and females were found to be statistically significant, as females receive more
Next, respondents were asked how often they use the sales and promotions
disseminated through Facebook, and the results of the t-test for independent samples
which tested for differences between genders on this variable can also be found in Table
11. The Levenes Test for Equality found a calculated p-value of .002, which meant the
variances were not equal. The t-test revealed a calculated p-value of .001, which was less
than the desired significance level of .05. As a result, the differences between genders
The same process was repeated to determine whether differences between genders
were evident through the usage of social media network Twitter. The first question focused
on how often individuals checked their accounts on these sites, and t-tests for independent
samples were again used to examine differences between genders. As shown in Table 11, a
and a calculated p-value of .723 was discovered. As a result, the variances were
determined to be equal, and the top row of t-test output was used in the analysis. The
calculated p-value was .794, which was larger than the desired significance level of .
Next, the relationship between genders and number of Twitter followers was
tested using a t-test for independent samples. The results of this test can be found in
Table 11. When testing for equality of variance, a calculated p-value of .930 was
discovered; as a result, the variances were determined to be equal. The subsequent t-test
revealed a calculated p-value of .965. Because this p-value was larger than the desired
significance level of .05, the differences between males and females in terms of number
The third Twitter variable which was tested was the number of companies that
males and females follow on this social networking site. A t-test for independent samples
was again employed to test for differences, and the results of this test can be found in
Table 11. The initial Levenes Test for Equality of Variance found a calculated p-value
of .483, meaning the variances were equal. Consequently, the t-test found a p-value of
.226. Because this is larger than .05 (the desired level of significance), no statistically
significant differences were found between males and females in terms of the number of
Respondents were also asked how often they shop with the companies they follow on
Twitter, and a t-test for independent samples was used to test for differences between genders
on this variable. The Levenes Test for Equality of Variance found a calculated
42
p-value of .864, which indicated the variances were equal. Consequently, the calculated
p-value in the t-test was determined to be .994. Because this value was greater than the
desired significance level of .05, the differences in means between males and females
The next variable that was tested for gender differences was how often
these results can be found in Table 11. The initial Levenes Test for Equality of Variance
revealed the variances were equal, as the calculated p-value of .912 was larger than .05.
For the t-test, the calculated significance level was found to be .988, which was larger
than the desired level of .05. As a result, the frequency of which individuals receive
correspondence from companies they follow on Twitter does not vary by gender.
The next variable which was tested for differences between genders was how
often respondents receive sales information from companies they follow on Twitter. The
results of this analysis can be found in Table 11. A Levenes Test for Equality of
Variance revealed a p-value of .195, which was larger than .05. As a result, the variances
between the two groups were determined to be equal. A p-value of .586 was calculated
for the t-test of independent samples, and this value was larger than the desired
differences were found between genders in terms of how often they receive sales
The last variable which was tested was how often individuals use the discounts
and promotions disseminated to them through Twitter. This variable was then tested for
differences between genders. The results of this test can be found in Table 11. A
43
Levenes Test for Equality found a p-value of .655, meaning the variances between the
two groups were assumed to be equal. The t-test between the two means revealed a
calculated p-value of .835. Because this figure was larger than the desired .05, the
differences in means between the two groups were not found to be statistically
significant. As a result, unlike Facebook, where females use the promotions more often,
males and females use promotional items on Twitter at the same rate.
Research Question #3
frequency. To investigate this, the following research question and hypotheses were
established:
This question was analyzed by calculating the strength of the relationship between
usage frequency of social media sites and four separate factors: (1) how often individuals
shop at companies they have befriended on these sites, (2) how often individuals receive
general correspondence, (3) how often individuals receive sales information from companies
on these sites, and (4) how often they use sales information received on these sites. Because
correlations were used to test the strength of these relationships. These correlations were
tested at a significance level of .05. The results of these tests for variables relating to
Facebook can be found in Table 12, while the results of these tests relating to variables
Facebook
To first test this research question, respondents Facebook usage frequency
(survey question three) was related to how many companies they have liked on Facebook
(survey question five). The results of this test can be found in Table 12.
Table 12
Correlations
Fb Fb Fb Fb Fb Fb Fb
check friends likes shopfrequ correspondence salesinfo salesuse
Fb Pearson 1.000 .122 .278** .207** .262** .272** .137*
check Correlation
Sig (2- .054 .000 .001 .000 .000 .036
tailed)
N 249 248 241 236 237 237 236
was found to be weak. Based on these results, individuals who more frequently use
To test the third research question, the respondents Facebook usage frequency
(survey question three) was related to how often individuals shop at companies they have
correlation of .207, and this correlation was found to be significant at a .05 level of
45
significance. Thus, the relationship between respondents Facebook usage frequency and
how often they shop at companies they have befriended was found to be weak. Thus,
individuals who check Facebook frequently are slightly more likely to shop at stores they
Next, respondents Facebook usage (survey question three) was related to how
often they receive correspondence from these companies they have befriended on the site
(survey question eight). A Pearson correlation revealed a correlation of .262, which was
definite, but small. Thus, individuals who frequently use Facebook are not more likely to
Respondents Facebook usage (survey question three) was also related to how
often they receive information about sales, specials, or coupons from companies on this
social network (survey question nine). A Pearson correlation was conducted at a .05
level of significance, and a correlation coefficient of .272 was discovered. While this
correlated relationship. Thus, respondents who use Facebook more frequently are not
Finally, respondents Facebook usage (survey question three) was related to how
often they utilize the sales and specials provided from companies they had befriended on
the site (survey question ten). A Pearson correlation indicated a correlation coefficient of
.137, and this relationship was found to be statistically significant at a .05 level
of significance.
46
of relationships between usage frequency and several other variables. These results can
Table 13
Correlations
Twitter Twitter Twitter Twitter Twitter Twitter Twitter
check friends likes shopfrequ correspondence salesinfo salesuse
Twitter Pearson 1.000 .483** .379** .465** .461** .479** .444**
check Correlation
Sig (2- .000 .000 .000 .000 .000 .000
tailed)
N 226 176 213 211 209 211 209
First, the relationship between Twitter usage frequency (survey question 11) and
the number of companies that individuals followed on Twitter (survey question 13) was
examined. Because the scales were interval, a Pearson correlation was used at a
significance level of .05. A correlation coefficient of .379 was calculated, and this was
definite, but small. Thus, individuals who use Twitter more often are not more likely to
Next, the relationship between respondents Twitter usage (survey question 11)
and shopping frequency at stores they were following (survey question 15) was tested. A
Pearson correlation revealed a correlation coefficient of .465, and this figure was
substantial, meaning individuals who frequently use Twitter may be more likely to shop
A Pearson correlation was also used to evaluate the relationship between Twitter
usage frequency (survey question 11) and how often individuals receive general
correspondence from these companies (survey question 16). This test indicated a
correlation coefficient of .461, which was statistically significant at a level of .05. This
Twitter may be more likely to receive information from companies through the
networking site.
Next, the relationship between how often individuals use Twitter (survey question
11) and how often they receive information about sales, specials, or coupons through the site
(survey question 17) was examined. Because these scales were interval, a Pearson correlation
at a significance level of .05 was used to analyze this relationship. The Pearson test indicated
05. Therefore, individuals who use Twitter more frequently may be more likely to receive
Finally, the relationship between how often respondents use Twitter (survey
question 11) and how often they utilize sales, coupons or promotions from the site
(survey question 18) was evaluated using a Pearson correlation at a significance level of
.05. This analysis revealed a correlation coefficient of .444 which was statistically
significant. However, this relationship was still classified as moderate. Thus, individuals
who use Twitter more frequently are not more likely to take advantage of the sales
At first glance, these results may seem inconsistent; however, further analysis
explains these results. The first correlations for each social networking site showed that
heavy Facebook and Twitter users were not more likely to befriend companies;
conversely, these respondents may be using these sites to network with other individuals
rather than companies. Because these heavy Facebook and Twitter users are not
necessarily befriending companies, they were less likely to shop at companies they
connected with on these sites, receive general correspondence, and obtain information
on sales. Based on this analysis, companies should not segment their markets based on
usage patterns; rather, they should target the users who are more likely to disseminate
information.
Regressions
the study on the dependent variable of sales usage frequency, multiple regression
analyses were conducted. The results of these analyses can be found in Tables 14
through 19. Tables 14, 15, 16 relate to the effects of the independent variables on sales
usage frequency of deals provided on Facebook, while Tables 17, 18, and 19 relate to
the effects if the independent variables on sales usage frequency of deals provided
through Twitter. Each of these analyses was conducted at a significance level of .05.
49
Table 14
Model Summary
Std. Error of the
Model R R Square Adjusted R Square Estimate
1 a .577 .546 .787
.759
a. Predictors: (Constant), ethnicgroup, fbsalesinfo, fbfriends, studenttype,
gender, fbcheck, maritalstatus, workstatus, fblikes, fbshopfrequ, classrank,
fbcorrespondence, age
individuals used sales provided to them through Facebook, the following independent
variables were analyzed: gender, class rank, student type, work status, marital status,
promotions received from companies on Facebook, age, and ethnic group. As shown in Table
14, the R Square value was .577, meaning that 57.7% of the variation in sales usage can be
explained by the variables analyzed in this analysis. The Adjusted R-Square, which accounts
Table 15
ANOVA Results for Multiple Regression Analysis of Variables on Sales Usage Frequency
of Deals on Facebook
b
ANOVA
Model Sum of Squares df Mean Square F Sig.
1 Regression 148.713 13 11.439 18.455 a
.000
Residual 109.098 176 .620
Total 257.811 189
a. Predictors: (Constant), ethnicgroup, fbsalesinfo, fbfriends, studenttype, gender, fbcheck, maritalstatus, workstatus,
fblikes, fbshopfrequ, classrank, fbcorrespondence, age
b. Dependent Variable: fbsalesuse
Table 16
a
Coefficients
Standardized
Unstandardized Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) -.206 .536 -.384 .701
gender .223 .124 .091 1.802 .073
classrank .027 .058 .037 .457 .649
studenttype -.075 .188 -.024 -.399 .690
workstatus .021 .093 .015 .232 .817
maritalstatus .063 .195 .020 .324 .746
fbcheck -.132 .063 -.110 -2.090 .038
fbfriends .041 .037 .059 1.131 .260
fblikes .082 .039 .132 2.127 .035
fbshopfrequ .254 .061 .269 4.176 .000
fbcorrespondence .080 .069 .087 1.169 .244
fbsalesinfo .327 .066 .381 4.927 .000
age .061 .049 .116 1.245 .215
ethnicgroup .069 .055 .063 1.245 .215
F(13, 176)=18.46, p <.000. Because the calculated significance level (.000) was less than
the desired level of significance (.05), this relationship was deemed to be linear. The
which were significant, the calculated levels of significance for each variable were
considered; if these levels of significance were less than the desired significance level of
.05, then the coefficient was deemed to be significant. In this case, the coefficients of
which information on sales is received from companies on Facebook (p < .000) were all
more frequently at these companies, and receive more sales information, they are more
likely to take advantage of these promotional offers. Interestingly, this analysis also
seems to indicate that as individuals check their Facebook accounts more often, they are
less likely to take advantage of the sales and promotional offers they receive through this
medium.
Next, the same test was conducted to determine which variables influenced the
frequency at which individuals used sales provided to them through Twitter. To this extent,
the following independent variables were analyzed: gender, class rank, student type, work
status, marital status, frequency of checking Twitter, number of Twitter followers, number of
age, and ethnic group. An analysis of variance (ANOVA) test was used, and the model
Table 17
Model Summary
Std. Error of the
Model R R Square Adjusted R Square Estimate
1 a .768 .745 .504
.877
a. Predictors: (Constant), twittersalesinfo, age, gender, ethnicgroup, twittercheck,
studenttype, maritalstatus, workstatus, twittercompanies, twitterfollowers,
classrank, twittershopfreq, twittercorrespondence
As shown in Table 17, the R Square value was .768, meaning that 76.8% of variance in
the dependent variable (sales usage frequency of deals on twitter) was explained by the
Table 18
Table 19
a
Coefficients
Standardized
Unstandardized Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) .215 .345 .624 .533
gender -.026 .090 -.013 -.291 .771
classrank .021 .043 .033 .483 .630
studenttype .191 .149 .064 1.287 .200
workstatus -.104 .069 -.080 -1.507 .134
maritalstatus -.124 .158 -.040 -.783 .435
age -.019 .038 -.038 -.484 .629
ethnicgroup .046 .043 .048 1.075 .284
twittercheck .036 .031 .059 1.164 .247
twitterfollowers -.031 .041 -.042 -.739 .461
twittercompanies -.091 .044 -.115 -2.088 .039
twittershopfreq .111 .079 .118 1.417 .159
twittercorrespondence .446 .111 .433 4.010 .000
twittersalesinfo .369 .112 .397 3.291 .001
a. Dependent Variable: twittersalesuse
Table 18 shows the results of the subsequent analysis of variance (ANOVA) test,
which was F(13, 129)=32.94, p <.000. As shown in this table, the calculated significance
level (p < .000) was less than the desired level of significance (p <.05); therefore, this
relationship was deemed to be linear. The results found in Table 19 indicate which
coefficients were significant. Again, the calculated significance level for each coefficient
was compared to the desired level of significance of .05. If the calculated level of
significance was lower than .05, then that variable was deemed to be significant. In this
correspondence was received from companies on Twitter (p < .000), and frequency at
which information on sales is received from companies on Twitter (pL.001) were deemed
54
and sales information are more likely to take advantage of these sales. Surprisingly, as
individuals follow more companies on Twitter, they are less likely to take advantage of
sales offers. At first, this may seem counter-intuitive; however, individuals who follow a
large number of companies may reach information overload and may not be able to take
Chapter 5: Discussion
Though this study focused on only on college students, the results of this study
are consistent with previous studies on the role of social media on consumption
patterns. In this section, the results of the study will be discussed and will be analyzed
Research Question #1
websites like Facebook and Twitter affect consumption patterns. To examine this
research question, the following research question and hypotheses were established:
on consumption patterns.
on consumption patterns.
frequency of the sales, specials, or coupons received from these sites. For each variable,
two sets of data were collected: one for Facebook and one for Twitter. The relationships
between variables were then tested using Pearson correlations at a significance level of
.05.
56
The first variable which was analyzed was number of Facebook friends, and this
variable was found to be related significantly to one other variable: number of companies
liked on Facebook (.140). This statistic indicates that individuals who have more friends
on this social networking site are more likely to like companies; however, this
relationship is very weak and almost negligible. This relationship was found to be
Thus, individuals who follow a larger number of other individuals are more likely to
number of Facebook friends and other variables, several substantial relationships were
found between number of individuals followed on Twitter and other variables. For
frequency at stores followed on Twitter was found to be .505, meaning individuals who
build larger social networks on the site are more likely to shop at stores they have
connected to on Twitter. Likewise, individuals with larger social networks are more
likely to receive correspondence from companies (.458), receive sales information (.537),
The results summarized above suggest that individuals may be using Twitter as a
note that this variable related to the number of individuals that respondents were
following, rather than companies. Given these moderate relationships and the low
57
discovered. First, a substantial relationship (.487) was found between the variables of
relationship was even stronger on Twitter (.523). These results indicate individuals are
likely to shop at companies they have followed on social media sites, as well as connect
The findings of the study also found individuals who follow or friend a larger
number of companies are more likely to receive correspondence from these companies,
as the correlations for these two variables were .389 for Facebook and .554 for Twitter.
These results indicate that some companies are already using these social networking
with larger numbers of companies on these sites were likely to receive more sales
information than individuals who did not follow a larger number of companies. Again,
this correlation was stronger on Twitter (.582) than it was for Facebook (.419). These
results seem to show that many companies are already adopting Twitter as a
communication channel with customers. As shown by Hyllegard, Ogle, Yan, and Reitz
(2011), customers may gain additional benefits from receiving this information through
social media: having information about companies or sales may make them market
mavens who can share this information with others. These motivations then relate to the
58
followed on social media sites and the usage of sales information (the correlation was
.444 on Facebook and .484 on Twitter). Thus, individuals who are seeking out
companies on social media networks are then leveraging the information they receive
to gain additional benefits from the relationship. These findings are consistent with
Foregger (2008), who applied the uses and gratifications model to Facebook. Foregger
(2008) found one of the reasons individuals used Facebook was for
Shopping Frequency
The third variable which was analyzed was how frequently individuals shopped at
companies they had liked or followed on social networking sites. The results of Pearson
correlations on this variable indicated moderate to high correlations between the other
variables. For instance, individuals who shopped more often at companies they had
information from these companies. These results indicate one of two options: (1)
individuals are seeking out these companies online and are connecting with them to
social media sites and this information is motivating them to shop more at these stores.
Either result is good news for companies that are seeking new channels of
communication with customers. These results were consistent when analyzing the
relationship between shopping frequency and sales information. A correlation of .565 for
59
Facebook and .855 on Twitter indicates that companies are using these social media sites to
disseminate information about sales and special offers. Because Pearson correlations do not
show causal effects, either individuals who visit the stores are more likely to receive sales
redemption of sales offers. The correlations between these two variables were moderate
(.612 for Facebook) to strong (.792 for Twitter), thus indicating a relationship between
the two variables. Consequently, these social media strategies are having the desired
effect: individuals are visiting the stores to take advantage of the offers provided through
Frequency of Correspondence
companies are sharing through social media. As shown by Barnes (2008), many
individuals prefer to research companies and products through social media prior to
and Twitter (a correlation of .918), it appears much of the correspondence which is being
sent is dedicated to informing customers about sales and special offers. In addition,
of correspondence from companies are more likely to redeem offers from these
Finally, respondents were asked how often they receive the sales promotions
provided through social media. The results of this variable were correlated with
responses to how often respondents redeem sales offers. In both cases, the correlations
were significant, as the correlation between these variables for Facebook was .628 and
the correlation between these two variables for Twitter was .868. These findings show
College students who receive greater amounts of information on sales offers are more
likely to redeem these sales. In such instances, all parties benefit. Companies increase
their market penetration through higher sales, while more effectively targeting only
interested customers and reducing wasted coverage. Customers, meanwhile, receive only
relevant marketing materials from these firms, and they can opt-out at any time by no
longer friending or following the companies. These findings are consistent with the
findings of Hu, Liu, and Zhang (2008), who showed that, though influence may be
consumers.
Research Question #2
conducted to compare the mean ratings between males and females on 14 different
often correspondence was received from companies liked on Facebook, how often sales
information was received from companies liked on Facebook, how often sales
promotions received through Facebook were used, usage frequency of Twitter, number of
stores followed on Twitter, how often correspondence was received from companies
followed on Twitter, how often sales information was received from companies followed
on Twitter, and how often sales promotions received through Twitter were used.
The first variable which was analyzed was the usage frequency of social media
differences between males and females in terms of usage frequency of Facebook and
Twitter. These results are consistent with past research studies conducted by Budden,
Anthony, Budden, and Jones (2007) and Gerlich, Browning, and Westermann (2010),
who found that statistically significant differences did not exist between genders in terms
62
of social media usage. However, the results of the present study are inconsistent with the
findings of Pempek, Yermolayeva, and Calvert (2009), who found that females spent
more time on social media sites. As a result, companies can effectively reach members of
both genders by creating promotional strategies based around social media sites like
friends/followers on Facebook and Twitter. Again, the independent t-tests did not reveal
statistically significant differences between males and females in terms of the number or
friends of followers on these social networking sites. These results are inconsistent with
the findings of Pempek, et al. (2009), who found that females were likely to have more
Facebook friends than males. These differences may exist for several reasons. First, the
sample in the present study was taken from only one campus, and social media
behaviors at this campus may differ from those present elsewhere. In addition, the study
conducted by Pempek et al. was conducted three years ago, and male social networking
The third variable which was analyzed was the number of companies that were
differences were discovered through independent t-tests. These results are thus similar to
the results found when comparing the number of friends or followers on these sites, and
they seem to indicate that businesses may be able to reach males and females equally
males or females were more likely to shop at stores they had connected with through
63
social media. Independent t-tests revealed that no significant differences existed between
genders; thus, males and females were both equally likely to shop at stores they had
males and females in terms of the amount of correspondence they received from
significant differences did not exist between genders. This result is not surprising; if no
differences exist between the number of companies followed by males and females,
then both genders should receive the same amount of correspondence from companies.
between gender and social media outlets when analyzing sales information. Independent
t-tests were conducted to test whether males and females received different amounts of
sales information from companies through social media. Females were found to receive a
receiving more sales information than males on Facebook can be explained in two ways.
First, since no statistical differences existed between males and females in terms of
companies friended or liked, companies may be targeting females through social media;
consequently, these companies are sending more sales information to this segment.
Alternatively, females may be more attune to these deals and thus may be more apt to
notice when companies are providing these offers. The fact that no statistically significant
received on Twitter might be due to the newness of this social media site relative to
Facebook.
The last variable which was analyzed was the usage of sales information between
genders. Independent t-tests found that females were more likely to redeem sales offers
provided through Facebook than males; however, no differences existed between males
and females on this variable regarding Twitter. These results are logical based on other
findings in the present study: if women receive sales information more often (as they do
on Facebook), then the odds are greater that they will take advantage of these deals.
However, if both genders receive the same amount of sales information (as they do on
Twitter), then they are equally likely to redeem these deals. The study also revealed that
some individuals are taking advantage of these sales offers and deals, thus supporting the
findings of Hennig-Thurau et al. (2004) and Hu, Liu, and Zhang (2008) who found that
electronic word-of-mouth media like social networks can be used to shape opinions and
improve sales.
Research Question #3
frequency. To investigate this research question, the following research question and
between usage frequency of social media sites and four separate factors: how often
individuals shop at companies they have befriended on these sites, how often individuals
receive general correspondence, how often individuals receive sales information from
companies on these sites, and how often they use sales information received on these
sites. Because the scales used on the survey instrument were interval-level, Pearson
The first relationship which was examined was between usage frequency and
correlation conducted on these two variables for Facebook resulted in a .207 correlation
coefficient, while a Pearson correlation for these two variables for Twitter resulted in a
significant, though the Facebook correlation was deemed to be definite but small and the
Twitter correlation was substantial. These results indicate that individuals who use social
media more often tend to shop more often at companies they have friended on the sites.
The second relationship which was analyzed was between the variables of usage
frequency and correspondence frequency. Pearson correlation tests conducted on the two
Facebook variables indicated a .262 correlation coefficient, while the Pearson correlation
for the two Twitter variables showed a .461 correlation coefficient. Thus, the relationship
66
between these two variables for Facebook was deemed low and small, while the
relationship between these two variables for Twitter was deemed moderate and
substantial. As a result, individuals who use social media more often are likely to receive
Next, the relationship between how often individuals use social media and how
often they receive sales information from companies on these sites was analyzed using
Pearson correlations. These tests indicated a small .272 correlation coefficient for these
variables for Facebook, while a Pearson correlation for these two variables for Twitter
resulted in a substantial .479 correlation coefficient. Thus, individuals who use social
media more frequently are more likely to receive promotional material from companies
Finally, the relationship between how often individuals use social media and how
often they redeem sales and offers was analyzed. A Pearson correlation conducted on
these two variables for Facebook resulted in a negligible .137 correlation coefficient,
while a Pearson correlation for these two variables for Twitter resulted in a substantial
.444 correlation coefficient. These results indicate that individuals who use social media
more often are more likely to take advantage of sales offers; however, this relationship is
These last two results are consistent with the findings of Foregger (2008), who
applied the uses and gratifications model to Facebook. In that study, Foregger (2008)
found nine factors which influenced Facebook usage: pass time, connection, sexual
comparison, channel use, and networking. The results of the present study relate to
67
Foreggers (2008) accumulation factor, which relates to the use of the social networking
site to buy and sell items. Foreggers (2008) results indicate that the ability to buy
products from the site was a factor in the usage of this social networking site, and the
results of the present study indicate that individuals who use social media more often are
Summary
Businesses are already using social networking websites to reach customers. The
results of this study indicate that students receive and use the promotions they receive
through social media sites Facebook and Twitter. However, it was beyond the scope of
Regardless, the arrangement is beneficial to all parties. Businesses that use social media
could use both Facebook and Twitter. Moreover, because no statistically significant
differences were discovered between genders in terms of social media usage, companies
are able to utilize both Facebook and Twitter to reach all customers. Nevertheless,
emphasize their female clientele should consider Facebook rather than Twitter. Overall,
college students who use social media more frequently receive more correspondence
from companies, and these students then use the promotions. Therefore, businesses
Limitations
Several limitations were present in the study. First, the data were collected via
convenience sample. Although convenience samples provide time advantages, they may
be biased and may not represent the entire population. Also, the survey relied on self-
reporting. Based on the respondents perceptions, they may not have accurately reported
their behavior. Additionally, students may have taken the survey multiple times. Because
the IP addresses were not recorded, there is not a way to check this information.
Furthermore, almost half of the respondents (49.8%) were 18-21 years old. Because they
are younger, they may be more apt to check their social media accounts, whereas other
ages may have other responsibilities, such as full-time employment and families. In
addition, the study was conducted at a single institution; although these results may not
seem generalizable beyond this college, the institution has indicators that it is not
Heuristics
Based on the findings of the study, several additional studies could be developed.
First, the study could be replicated at several different institutions to determine whether
the results at the present testing site were typical or were atypical of the general
population of college students. Though it was beyond the scope of this study, further
analysis could be performed to test for differences between demographics. For example,
patterns of students from different majors, ages, or ethnic groups could also be tested.
Future research could also seek to determine whether the effectiveness of word-
of-mouth communications varies based on the industry of the sender. For example,
continues to advance and new social networking platforms are adopted, the present study
could be expanded to include these additional channels, while also tracking the general
Appendixes
71
Appendix A:
Questionnaire
How many types of social networking accounts do you maintain?
___ I do not have any social networking accounts.
___ 1
___ 2
___ 3
___ 4
___ 5
___ 6 or more
Very frequently
5
Very frequently
5
How often do you shop at the stores you have befriended/liked on Facebook? (Circle)
Never Very frequently
1 2 3 4 5
How often do you receive correspondence from the companies you befriended/liked on
Facebook?
(Circle)
Never Very frequently
1 2 3 4 5
How often do you receive information about sales, specials or coupons from the
companies you befriended/liked on Facebook? (Circle)
Never
1 2 3 4
How often do you utilize the sales, specials or coupons from the companies you
befriended/liked on Facebook? (Circle)
Never
1 2 3 4
Very
freq
5
Very
freq
5
Very
freq
5
How many followers do you have on Twitter?
___ More than 200
___ 150-199
___ 100-149
___ 50-99
___ 1-49
___ I do not have a Twitter.
How many companies/businesses do you follow?
___ None
___ 1-5
___ 6-10
___ 11-15
___ 16-20
___ More than 20
How often do you shop at the stores you follow on Twitter? (Circle)
Never Very frequently
1 2 3 4 5
How often do you receive correspondence from the companies you follow on Twitter?
(Circle)
Never
1 2 3 4
How often do you receive information about sales, specials or coupons from the
companies you follow on Twitter? (Circle)
Never
1 2 3 4
How often do you utilized the sales, specials or coupons from the companies you
follow on Twitter? (Circle)
Never
1 2 3 4
74
Are you
___ Male
___ Female
Marital status:
_____ Single _____ Married/Partner
75
Appendix B:
Research Questions
To understand the effects of these websites on the consumption patterns of college
students, the following three research questions were established:
Q1. Do word-of-mouth recommendations on social media websites
affect consumption patterns?
Q2. Does the effectiveness of word-of-mouth recommendations on social media
differ between genders?
Q3. Does the effectiveness of word-of-mouth recommendations on social media
websites differ based on social media usage frequency?
Survey Instrument
To answer these questions, a survey was created and distributed electronically
through the website SurveyMonkey. Depending on student responses, the survey
instrument contained between 10 and 28 questions. First, respondents were asked how
many social networking accounts they maintained. Next, respondents were given a list of
common social networking sites, and were asked to identify which of these they used.
Using skip logic technology, students who confirmed they owned Facebook
accounts were asked a series of questions regarding their usage of this social networking
site. First, respondents were asked to indicate their frequency of usage on a 5-point
Likert-type scale. Next, respondents were asked how many friends they had on the
site, as well as how many businesses they had liked on the site. Respondents were also
encouraged to list the businesses they had liked. Finally, students were asked to indicate,
using a 5-point, Likert-type scale how often the shopped at friended stores, how often
they received correspondence from these stores, how often they received special sales
offers from these stores, and how often they redeemed these special offers. This process
was repeated for students who indicated they had maintained a Twitter account.
Finally, basic demographic information was collected from the students. To this
extent, they were asked to report their gender, class rank, major, age, student status,
employment status, marital status, and ethnic group. This survey was then approved by
the committee and schools IRB before being distributed to students on campus.
Methodology
The electronic survey was completed by 275 students at a small Midwestern
college. Participants were chosen via a convenience sample, as they were provided with
77
the surveys web address and asked to visit the site to complete the survey. This survey
was distributed for approximately one month, from mid-February until mid-March 2012.
Results
In order to analyze this research question, the relationships between several
variables were explored for both Facebook and Twitter: number of companies liked;
shopping frequency at liked stores; frequency of correspondence; frequency of sales,
specials or coupons; and usage frequency of the sales, specials or coupons received from
these sites. Because the scales used represented interval-level data, Pearson correlations
were used at a significance level of .05. The tests found that students who like more
companies on Facebook receive more correspondence from these companies as well as
are likely to utilize the promotions while shopping there. Furthermore, when students
follow companies on Twitter, they receive information from the company but only
slightly more correspondence than those who do not follow the business. However, those
individuals who receive correspondence from companies are more likely to shop there as
well as receive special deals.
The second research question relating to differences between genders was
examined using t-tests for independent samples to determine whether the differences in
means were statistically significant at a significance level of .05. These tests revealed
that women are more likely to receive information about sales and promotions from
companies on Facebook, and they are more likely to take advantage of these offers.
However, no statistically significant differences were discovered between males and
females in terms of the role of Twitter in shaping their consumption patterns.
The third research question was tested by examining the relationship between
usage frequency was tested against six other variables: number of friends, number of
companies liked/followed, how often respondents shopped at liked/followed companies,
how often respondents received correspondence from stores liked/followed, how often
respondents received sales or promotions from companies liked/followed, and how often
respondents used these offers. These relationships were tested using Pearson correlations
at a significance level of .05. Individuals who frequently use Facebook are not more
likely to receive correspondence from a company that has been befriended; however,
people are more likely to shop at stores they have networked with on the site. Similar
results were found with Twitter with one difference: individuals who use Twitter more
frequently are more likely to receive additional information on sales and promotions.
However, the more companies a person likes/follows and the more correspondence
received, the less likely the individual is to utilize the promotions.
Further analysis between variables was conducted using analysis of variance tests
(ANOVAs) at a significance level of .05.
Limitations
Three major limitations existed in the current study:
Data was collected via a convenience sample.
Data was collected at a single institution.
Data relied on self-reporting.
78
Heuristics
Several areas of future research could be explored, based on the findings of this study:
Replicating the study at another institution.
Further analysis could be performed to test for differences between
demographics, such as class rank, major, ages, or ethnic groups.
The effectiveness of social media promotions across several different industries
could be examined.
Over time, the effectiveness of new social media sites could be tested relative to
the enduring strength of current social media sites.
Conclusion
Businesses are already using social networking websites to reach customers. The
results of this study indicate that students receive and use the promotions they receive
through social media sites Facebook and Twitter. The arrangement is beneficial to all
parties. Moreover, because no statistically significant differences were discovered
between genders in terms of social media usage, companies are able to utilize both
Facebook and Twitter to reach all customers. Nevertheless, females receive or notice
more correspondence on Facebook; therefore, businesses that emphasize their female
clientele should consider Facebook rather than Twitter. Overall, college students who
use social media more frequently receive more correspondence from companies, and
these students then use the promotions. Therefore, businesses should use Facebook and
Twitter to obtain the patronage of students but be wary of overloading them with too
much information.
79
Appendix C:
Works Cited
Barnes, N. G. (2008). Exploring the link between customer care and brand reputation in
the age of social media. Unpublished paper, New Communications Forum,
Sonoma County, CA.
Budden, C.B,, Foster Anthony, J., Budden, M. C., & Jones, M. A. Managing the
evolution of a revolution: Marketing implications of internet media usage among
college students. College Teaching Methods & Styles Journal, 3(3), 5-10.
Retrieved October 16, 2011
Casteleyn, J., Mottart, A., & Rutten, K. (2009). How to use Facebook in your market
research. International Journal of Market Research, 51(4), 439-447.
Eldon. (2011). Facebook stats. In Inside Network. Retrieved October 20, 2011
Ellison, N. B., Steinfield, C., & Lampe, C. (2007). The benefits of Facebook friends:
Social capital and college students' use of online social network sites. Journal
of Computer-Mediated Communication, 12(4), 11431168.
Gerlich, R. N., Browning, L., & Westermann, L. (2010, November). The social media
affinity scale: implications for education. Contemporary Issues in Education
Research, 3(11), 35-41. Retrieved November 28, 2011.
Glazer, E. (2011, February 7). Fund firms cautiously tweet their way into a new world.
The Wall Street Journal. Retrieved October 16, 2011.
Hennig-Thurau, T., Gwinner, K. P., Walsh, G., & Gremler, D. D. (2004). Electronic
word-of-mouth via consumer-opinion platforms: What motivates consumers to
articulate themselves on the internet? Journal of Interactive Marketing, 18(1),
38-52. Retrieved November 9, 2011.
Hu, N., Liu, L., & Zhang, J. J. (2008, March 11). Do online reviews affect product sales?
The role of reviewer characteristics and temporal effects. Information
Technology and Management, 9(3), 201-214. doi:10.1007/s10799-008-0041-2
Hyllegard, K. H., Ogle, J. P., Yan, R., & Reitz, A. R. (2011, September). An exploratory
study of college students' fanning behavior on Facebook. College Student
Journal, 45(3), 601-616. Retrieved October 16, 2011.
93
LaDuque, J. (2010, July). Get more leads, engage customers with social media. In IFA
International Franchise Association. Retrieved October 16, 2011.
Pempek, T. A., Yermolayeva, Y. A., & Calvert, S. L. (2009). College students' social
networking experiences on Facebook. Journal Of Applied Developmental
Psychology, 30(3), 227-238. doi:10.1016/j.appdev.2008.12.010
Ross, C., Orr, E. S., Sisic, M., Arseneault, J. M., Simmering, M. G., & Orr, R. (2009).
Personality and motivations associated with Facebook use. Computers In
Human Behavior, 25(2), 578-586. doi:10.1016/j.chb.2008.12.024
Smock, A., Ellison, N.B., Lampe, C. & Wohn, D.Y. (2011). Facebook as a Toolkit: A
Uses and Gratification Approach to Unbundling Feature Use. Computers in
Human Behavior.